Genomes of Abundant and Widespread Viruses from the Deep Ocean

The deep sea is a massive, largely oligotrophic ecosystem, stretched over nearly 65% of the planet's surface. Deep-sea planktonic communities are almost completely dependent upon organic carbon sinking from the productive surface, forming a vital component of global biogeochemical cycles. However, despite their importance, viruses from the deep ocean remain largely unknown. Here, we describe the first complete genomes of deep-sea viruses assembled from metagenomic fosmid libraries. "Candidatus Pelagibacter" (SAR11) phage HTVC010P and Puniceispirillum phage HMO-2011 are considered the most abundant cultured marine viruses known to date. Remarkably, some of the viruses described here recruited as many reads from deep waters as these viruses do in the photic zone, and, considering the gigantic scale of the bathypelagic habitat, these genomes provide information about what could be some of the most abundant viruses in the world at large. Their role in the viral shunt in the global ocean could be very significant. Despite the challenges encountered in inferring the identity of their hosts, we identified one virus predicted to infect members of the globally distributed SAR11 cluster. We also identified a number of putative proviruses from diverse taxa, including deltaproteobacteria, bacteroidetes, SAR11, and gammaproteobacteria. Moreover, our findings also indicate that lysogeny is the preferred mode of existence for deep-sea viruses inhabiting an energy-limited environment, in sharp contrast to the predominantly lytic lifestyle of their photic-zone counterparts. Some of the viruses show a widespread distribution, supporting the tenet "everything is everywhere" for the deep-ocean virome.

The deep sea is among the largest known habitats and a critical cog in biogeochemical cycling but remains underexplored in its microbiology. Even more than is the case for its prokaryotic community, our knowledge of its viral component has remained limited by the paucity of information provided by studies dependent upon short sequence fragments. In this work, we attempt to fill this existing gap by using a combination of classical fosmid libraries with next-generation sequencing and assembly to recover long viral genomic fragments. We have sequenced ca. 6,000 fosmids from two metagenomics libraries made from prokaryotic biomass from the deep Mediterranean Sea and recovered twenty-eight complete viral genomes, all of them novel and quite distinct from all previously described viral genomes. They are preferentially found in deeper waters and are widely distributed all over the oceans. To our knowledge, this is the first report on complete and cosmopolitan viral genomes from the bathypelagic habitat.